RESEARCH Open Access

Cervical small cell carcinoma frequentlypresented in multiple high risk HPVinfection and often associated with othertype of epithelial tumorsPeifeng Li†, Jing Ma†, Xiumin Zhang, Yong Guo, Yixiong Liu, Xia Li, Danhui Zhao and Zhe Wang*

Abstract

Background: Small cell carcinoma of the uterine cervix is a rare and highly malignant tumor, and its etiopathogenesisis strongly related to high-risk HPV infections.

Methods: The clinicopathological data of 30 cases of cervical primary small cell carcinoma were retrospectivelyanalyzed. In situ hybridization, polymerase chain reaction and reverse dot-blot hybridization were employed to detectHPV DNA in both small cell carcinoma and other coexisting epithelial tumors. Immunohistochemistry was used todetect the protein expression of p16 and p53.

Results: Amongst 30 patients with cervical primary small cell carcinoma, 15 patients simultaneously exhibited other typesof epithelial tumors, including squamous cell carcinoma, adenocarcinoma, squamous cell carcinoma in situ, andadenocarcinoma in situ. Most tumor cells infected with HPV presented integrated patterns in the nuclei by in situhybridization. HPV DNA was detected in every small cell carcinoma case (100%) by polymerase chain reaction and reversedot blot hybridization. 27 cases (90%) harbored type 18, and 15 (50%) displayed multiple HPV18 and 16 infections. Theprevalence of HPV 18 infection in small cell carcinoma was higher than in cervical squamous and glandular epithelialneoplasms (P = 0.002). However, similar infection rates of HPV 16 were detected in both tumors (P = 0.383). Both small cellcarcinoma and other types of epithelial tumors exhibited strong nuclear and cytoplasmic staining for p16 in all cases.Three cases of small cell carcinoma revealed completely negative p53 immunohistochemical expression in 15 cases ofcomposite tumors, which suggested TP53 nonsense mutation pattern. The pure small cell carcinoma of uterine cervix hadsimilar mutation or wild type pattern for TP53 compared with composite tumor (P = 0.224).

Conclusions: Cervical small cell carcinomas are often associated with squamous or glandular epithelial tumors, whichmight result from multiple HPV infections, especially HPV 16 infection. Multiple HPV infections were not correlated withtumor stage, size, lymphovascular invasion, lymph node metastasis, or prognosis. Furthermore, careful observation ofspecimens is very important in finding little proportion of small cell carcinoma in the composite lesions, specifically incervical biopsy specimens, in order to avoid the missed diagnosis of small cell carcinoma.

Keywords: Human papillomavirus, p16, p53, Small cell carcinoma, Uterine cervix

* Correspondence: zhwang@fmmu.edu.cn†Peifeng Li and Jing Ma contributed equally to this work.State Key Laboratory of Cancer Biology, Department of Pathology, XijingHospital and School of Basic Medicine, The Fourth Military MedicalUniversity, Changle West Road #169, Xi’an 710032, Shaan Xi Province, China

© The Author(s). 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, andreproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link tothe Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.

Li et al. Diagnostic Pathology (2018) 13:31 https://doi.org/10.1186/s13000-018-0709-9

BackgroundNumerous clinical and experimental studies have demon-strated a closed etiopathogenetic relationship between thedevelopment of cervical cancers (including squamous cellcarcinoma, adenocarcinoma, and small cell carcinoma) andhigh-risk human papillomavirus (HPV) infection [1]. It hasbeen implicated that squamous cell carcinoma and adeno-carcinoma are correlated to HPV 16 infection, whilst cer-vical small cell carcinoma is linked to HPV 18 [2, 3]. As anuncommon and highly malignant tumor, small cell carcin-oma of uterine cervix has similar morphological features totumors arising in the lung. The clinicopathological featuresof cervical small cell carcinoma and its relationship withHPV infection has been widely studied, including caseseries and larger retrospective population-based reports [2,4]. Although some studies report that cervical small cellcarcinoma is often associated with other cervical cancer orintraepithelial neoplasia [5], the clinicopathological featuresof cervical composite tumors, including small cell carcin-oma, were rarely characterized. In these studies, HPV infec-tion was detected using a variety of methods [3, 6] such asimmunohistochemistry, polymerase chain reaction (PCR),in situ hybridization (ISH), reverse dot blot hybridization(RDDH), and Southern blot hybridization. The inconsistentresults affected the understanding of the relationship be-tween HPV infection and cervical small cell carcinoma.Many studies revealed the preponderant infection of HPV18 in cervical small cell carcinoma [2], but multiple infec-tions were rarely reported.In this report, the clinicopathological features of 30 pa-

tients were analyzed, including 15 patients with pure cer-vical small cell carcinoma, and 15 cases with compositecervical tumors composed of small cell carcinoma andother types of epithelial tumors. The HPV infection in thesetumors was detected by ISH and PCR-RDDH, and the ex-pression of p16 and p53 proteins were examined using im-munohistochemistry. Our results demonstrate that cervicalsmall cell carcinoma is often associated with squamous orglandular epithelial tumors, and these tumor cells usuallyexhibit overexpression of p16. Small cell carcinomas andadenocarcinomas or squamous cell carcinoma in situshowed all completely negative p53 immunohistochemicalexpression in three of 15 composite tumors of uterine cer-vix. To the best of our knowledge, this is the first study toreport that multiple infections of HPV 18 and 16 developedin half of cervical small cell carcinomas. Furthermore, thisis the first report to make the distinction that HPV 18 in-fection is closely correlated to the occurrence of cervicalsmall cell carcinoma while HPV 16 infection is involved incervical squamous cell carcinoma or adenocarcinoma inthe co-existing tumor cases. Multiple infections of HPVsubtypes were not related to tumor stage, size, lymphovas-cular invasion, lymph node metastasis, or prognosis. Inaddition, our study also confirmed for the first time that

patients with composite tumors had similar HPV infectionsubtypes, clinicopathological features, and prognosis com-pared to patients with pure small cell carcinoma.

MethodsCase selectionThirty cases of cervical primary small cell carcinomawere retrieved from the 2009–2017 surgical pathologyarchives of the Xijing Hospital. These included 23 hys-terectomies, 1 conization, and 6 biopsies. To rule outthe possibility of metastatic small cell carcinoma fromthe lung or other organs, the clinical data were carefullyreviewed. Hematoxylin and eosin-stained slides of pri-mary cervical neoplasm were re-examined to confirmthe original diagnosis. Immunohistochemical staining forneuroendocrine markers, including Leu-19, synaptophy-sin, chromogranin, and neuron-specific enolase, was per-formed, and more than half of small cell carcinoma cellsshowed positive expression for two or more markers.The clinical records and follow-up of these patients wereexamined. This study received approval from the EthicsCommittee of the Xijing Hospital.

ISH analysisThree-micron sections containing small cell carcinomaand other types of epithelial tumors were examined forHPV DNA by ISH staining with INFORM® HPV IIIFamily 16 and 6 Probe (Ventana Medical Systems, Inc.,Tucson, AZ, USA) for high-risk HPV types (16, 18, 31,33, 35, 39, 45, 51, 52, 56, 58, and 66) and low-risk HPVtypes (6 and 11), respectively. The ISH assay was per-formed according to the manufacturer’s protocol usingthe Ventana BenchMark XT system (Ventana MedicalSystems, Inc.). Labeling was detected with the ISHiVIEW™ Blue Plus Detection Kit (Ventana Medical Sys-tems, Inc.). The positive and negative controls were car-ried out using HPV quality control slides provided byVentana Medical Systems, Inc. The HPV signals weredetected in the nuclei of tumor cells, and the signal pat-terns were categorized as either episomal staining pat-tern or punctuate staining pattern. The former presentsas a homogeneous globular navy-blue to black signal inthe entire nuclei of tumor cells, and the latter shows sin-gle or multiple sparsely distributed and dot-like navy-blue punctae in the nuclei of the tumor cells.

DNA extraction and HPV detection using PCR-RDDHEight-micron sections were prepared from formalin-fixed, paraffin-embedded (FFPE) tissues of cervicaltumor samples. Tumor tissues were dissected using ster-ile blades from slides and were collected into 1.5 mlEppendorf tubes. Amongst 15 specimens with compositecarcinomas, small cell carcinoma and other types of epi-thelial tumors were successfully isolated in seven

Li et al. Diagnostic Pathology (2018) 13:31 Page 2 of 10

specimens. 37 samples of tumor cells were collected, in-cluding 22 cases of small cell carcinoma, 8 cases ofmixed tumors, 2 cases of squamous cell carcinoma, 2cases of adenocarcinoma, 2 cases of squamous cell car-cinoma in situ, and 1 case of adenocarcinoma in situ.DNA extraction was carried out using QIAamp® DNAFFPE Tissue Kit (Cat No. 56404, QIAGEN GmbH, Hil-den, Germany) following the manufacturer’s guidelines.The samples were digested with proteinase K in a vol-ume of 200 μL at 56 °C overnight, and 20 μL of DNAaliquot was obtained finally. The quality of the genomicDNA extracted was tested by agarose gel electrophoresiswith ethidium bromide staining. For the PCR, 2 μL ofDNA aliquot was used, and broad-spectrum HPV DNAamplification was performed using primers for GP5+/6+(GP5+: 5’-TTTGTTACTGTGGTAGATACTAC-3′, GP6+: 5’-GAAAAATAAACTGTAAATCATATTC-3′) with atotal reaction volume of 50 μL. PCR amplification wasperformed in the following conditions: 50 °C for 15 min,95 °C for 10 min, 94 °C for 30 s, 46 °C for 90 s, 72 °C for30 s, 40 cycles, 72 °C for 5 min. DNA RDDH was per-formed with the HPV Genotyping Detection Kit (AsiaEnergy Biotechnology Co, Ltd., Shenzhen, China) tosimultaneously identify 23 HPV subtypes, including HPV6, 11, 16, 18, 31, 33, 35, 39, 42, 43, 45, 51, 52, 53, 56, 58,59, 66, 68, 73, 81, 82, and 83. PCR amplification prod-ucts were boiled 10 min to obtain single-stranded DNA.Samples were then added in low density gene patcheswith a total of 23 gene probes of different HPV subtypesand hybridization was performed at 51 °C for 3 h. Thegene patches were incubated in peroxidase solution for30 min, and then developed in color reagent (19 ml0.1 mol/L sodium citrate, 1 ml tetramethylbenzidine,and 2 μL 30% H2O2) for 60 min in the dark. HPV sub-types were determined by the positive point on the HPVgenotype profile on the membrane. HPV-positive andnegative controls were also included in everyexperiment.

ImmunohistochemistryImmunohistochemistry was performed on 3 μm thick tis-sue sections using the Ventana BenchMark XT system(Ventana Medical Systems, Inc.). P16 (clone: 6H12) andp53 (clone: DO-7) antibodies were purchased from MaxinCorp. (Fuzhou, China) and DAKO Corp. (Carpinteria,CA), respectively. Immunohistochemical staining wasconducted employing the Roche Ultraview DAB DetectionKit (Ventana Medical Systems, Inc.) following the manu-facturer’s instructions. The positive and negative controlslices were also run simultaneously. Nuclear/cytoplasmicstaining was considered positive for p16, and p53 proteinshowed nuclear expression. A tumor was recorded posi-tive for p16 if more than 50% of the tumor cells showedimmunoreactivity. IHC for p53 includes mutation pattern

and wild type pattern. Strong and diffuse nuclear stainingor complete negative with internal control is consideredto be mutation pattern, and focal and weak staining cor-relate with a wild type pattern of p53.

StatisticsStatistical software SPSS17.0 (SPSS, Inc., Chicago, IL,USA) was employed in this report. Pearson Χ2 test orFisher’s exact test was adopted for correlation analysis ofenumeration data. Wilcoxon rank sum test was used forcomparison of patients’ age and tumor size. P<0.05 wasconsidered as a statistically significant.

ResultsClinicopathological featuresAmong 30 patients with cervical primary small cell car-cinomas included in this study, 15 patients were diag-nosed with pure cervical small cell carcinoma. In theremaining 15 cases of composite tumors, 7 cases also ex-hibited invasive cervical squamous cell carcinoma, 4cases also showed cervical squamous cell carcinoma insitu, 3 cases also had cervical adenocarcinoma, and onepatient also displayed cervical adenocarcinoma in situ.The age of all 30 patients at diagnosis ranged from 31 to74 years (mean 46.4 years, median 41 years). The meanand median age of the 15 patients with cervical compos-ite tumors was 46.4 and 45 years, respectively. 22 pa-tients presented with abnormal vaginal bleeding orcontact bleeding. Two patients showed no clinical symp-toms, and the tumors were found following physicalexamination. Most of the tumors displayed exophyticgrowth with or without cervical erosion. The tumor sizesin the 24 hysterectomies and cervical conization speci-mens ranged from 1 to 6 cm (mean 3.0 cm, median2.5 cm). 16 tumors (66.7%) were FIGO stage IB, 3 stageIIA, 2 stage IA, 1 stage IIB, 1 stage IIIA, and 1 stageIVB. Accurate staging information was not acquired for6 cases of cervical biopsy specimens. 21 patients under-went radical hysterectomy with bilateral or partialadnexectomy and pelvic lymph node dissection. Two pa-tients received radical hysterectomy with pelvic lymphnode dissection, and one case underwent cervical coni-zation owing to the superficial invasion depth of smallcell carcinoma. Accurate surgical methods in 6 out-patients were not obtained in this study. 15 patientsunderwent postoperative chemotherapy combinedwith radiotherapy, and 6 patients received postopera-tive chemotherapy. Postoperative treatment strategieswere not obtained in 9 other patients. Follow-up dataof 16 cases was obtained: 7 patients died with a sur-vival time ranging from 5 to 24 months (median sur-vival time: 7 months). Survival time was more thantwo years in 4 out of 10 confirmed surviving patients,

Li et al. Diagnostic Pathology (2018) 13:31 Page 3 of 10

and two patients free of disease have undergone afollow-up at 67 and 88 months.Histologically, cervical primary small cell carcinomas

showed morphological features similar to those seen inthe pulmonary counterpart. Densely packed small tumorcells often formed a sheet-like diffuse growth pattern.Neuroendocrine growth patterns, such as orderly tubu-lar, trabecular, organoid, and nuclear palisading patterns,were less illustrated. Tumor cells showed round, ovoid,or spindled nuclei and scant cytoplasm. Nuclear chro-matin is finely granular, and nucleoli were absent or in-conspicuous. Numerous mitotic figures and extensivenecrosis were commonly observed. Furthermore, squa-mous and glandular epithelial neoplasms were observedin 15 cases. Most squamous cell carcinomas developedin the superficial parts of small cell carcinomas (Fig. 1a).However, in two cases, the squamous cell carcinomaintermingled with the small cell carcinoma (Fig. 1b). Inthree cases of cervical small cell carcinoma associatedwith adenocarcinoma, one case illustrated mixed smallcell carcinoma and adenocarcinoma (Fig. 1c), and twopatients revealed adenocarcinoma present at the periph-ery of the small cell carcinoma (Fig. 1d). Five cases ofcervical small cell carcinoma associated with carcinomain situ all illustrated adjacent relationships between smallcell carcinoma and squamous cell carcinoma in situ oradenocarcinoma in situ. Although more composite tu-mors were found in hysterectomy specimens than in bi-opsy specimens in this study, the discovery of compositelesions was not related to specimen type (P = 0.651). There

was no statistical difference in patients’ age, FIGO stage,lymph node metastasis, lymphovascular invasion, or prog-nosis between pure small cell carcinoma and compositetumors (Table 1, Fig. 2). Patients with cervical compositetumors had similar prognosis to patients with pure smallcell carcinoma (P = 0.716). The prognosis of these patientswas determined by the composition of small cell carcin-oma. Six cases (26.1%) had regional nodal metastasis, in-cluding 5 cases of metastatic small cell carcinoma and onecase of metastatic squamous cell carcinoma. One patientdisplayed small cell carcinoma involving vaginal stump,and ovarian metastatic small cell carcinoma was found inanother patient. In the 24 patients who underwent hyster-ectomies or cervical conization, lymphovascular invasionof small cell carcinoma was detected inside or within thetumor in 21 cases, of which two patients had simultaneouslymphovascular invasion of squamous cell carcinoma.Only one case presented with lymphovascular invasion ofsmall cell carcinoma in 6 cervical biopsy specimens. Lym-phovascular invasion was more likely to be found in hys-terectomy or conization specimens than in biopsyspecimens (P = 0.002).

HPV DNA detection and typingISH of tissue sections revealed that most specimenswere positive for the INFORM HPV III Family 16 Probein both small cell carcinomas and other types of epithe-lial tumors. 29 out of 30 cases of small cell carcinomasshowed positive staining, including 21 cases with onlyone navy-blue to black puncta (Fig. 3a), 6 cases with

Fig. 1 Cervical small cell carcinoma associated with other types of epithelial tumors (hematoxylin-eosin staining). a Squamous cell carcinoma (△)and squamous cell carcinoma in situ (◇) detected in the superficial parts of small cell carcinoma (☆) (× 40). b Squamous cell carcinoma (△)intermingled with small cell carcinoma (☆) (× 100). c Small cell carcinoma (☆) mixed with adenocarcinoma (※) (× 100). d Adenocarcinoma (※)present at the periphery of small cell carcinoma (☆) (× 100)

Li et al. Diagnostic Pathology (2018) 13:31 Page 4 of 10

multiple small blue puncta inside the nuclei (Fig. 3b),and 2 cases with multiple puncta-like staining to diffusesmall particles (Fig. 3c). These staining signals representviral integrative patterns in 27 cases. In some cases, thenuclear viral integrative pattern was so discrete thatmicroscopic examination at a higher magnification(40× objective) was required to judge nuclear epithelialcell localization, which represented very low viral copynumber. Two cases of small cell carcinoma showed both

episomal and integrative staining patterns. In 7 cases ofcervical squamous cell carcinoma, 5 cases illustratedpunctate signals in the nuclei of tumor cells (Fig. 3d) and2 cases displayed episomal staining patterns. Two out offour cases of squamous cell carcinoma in situ revealedepisomal staining patterns (Fig. 3e), and all tumor cells ofcervical adenocarcinoma and adenocarcinoma in situ il-lustrated punctate staining (Fig. 3f). Interestingly, one casedisplayed positive staining for small cell carcinoma withpunctate pattern but negative for squamous cell carcin-oma, and another was positive for cervical adenocarcin-oma in situ with punctate staining, but negative for smallcell carcinoma. The INFORM HPV III Family 6 Probewas employed to detect low-risk HPV genotypes, and nopositive tumor cells were observed in both small cell car-cinoma and other types of epithelial tumors.Thirty-seven specimens of cervical small cell carcin-

oma and other types of epithelial tumors were tested forHPV DNA by PCR-RDDH, and all were positive (Fig. 4).The prevalence of HPV 18 and 16 was 78.4% (29/37)and 64.9% (24/37), respectively. 48.6% (18/37) of speci-mens demonstrated multiple infections with variousHPV subtypes, and the predominant multiple infectionswere HPV 16 and 18 at a rate of 43.2% (16/37). The rareinfection of HPV 31, 33, and 43 was detected in threecases of small cell carcinoma with multiple HPV infec-tions. The prevalence of HPV 18 and 16 was 90.0%(27/30) and 60.0% (18/30) in the 30 cases of small cellcarcinoma, respectively. Multiple infections were foundin 17 cases of small cell carcinoma, among which coin-fection of HPV 18 and 16 was detected in 15 cases.

Table 1 Clinicopathological features of cervical small cell carcinoma with or without other types of epithelial tumors

Pure small cell carcinoma Small cell carcinoma associated with other epithelial tumors P value

Cases (n) 15 15

Specimen type 0.651

hysterectomy or cervical conization 11 13

biopsy 4 2

Age (years) 46.4 (34–68) 46.4 (31–74) 0.967

Size of tumors 3.4 (2–6) 2.5 (1–4.5) 0.105

FIGO stage 0.605

I 9 9

II 2 2

III 1

IV 1

Lymph node metastasis 1.000

Yes 3 3

No 8 9

Lymphovascular invasion 1.000

Yes 11 11

No 4 4

Fig. 2 Survival analysis of 17 patients with follow-up data. The patientswith composite cervical tumors had similar prognosis to patients withpure small cell carcinoma

Li et al. Diagnostic Pathology (2018) 13:31 Page 5 of 10

Multiple infections in composite tumors (73.3%) weremore common than in pure small cell carcinoma (46.7%), but there was no statistical difference (P = 0.136).The infection prevalence of HPV 16 and 18 was not sig-nificantly different between the composite tumor groupand the pure tumor group (P = 0.456 and 1.00, respect-ively). The HPV DNA subtypes in different tumor com-ponents were detected in seven cases of compositetumors, and the consistent HPV type was detected intwo cases with single infection. Four cases showed thattumor cells of small cell carcinoma were positive forHPV both 16 and 18, while other types of epithelial tu-mors, including adenocarcinoma, squamous cell carcin-oma, and squamous cell carcinoma in situ, onlydisplayed HPV 16 infection. The infection rate of HPV18 in small cell carcinoma was significantly higher thanin other types of epithelial tumors (P = 0.008). However,similar infection rates of HPV 16 were measured in bothgroups (P = 0.382, Table 2). In addition, multiple infec-tion with HPV subtypes was not related to tumor stage,size, lymphovascular invasion, lymph node metastasis, orprognosis (P = 0.187, 1.00, 1.00, 0.179, and 0.498,respectively).

Immunohistochemical expression of p16 and p53Strong and diffuse nuclear and cytoplasmic staining forp16 protein was noted in all cases of cervical small cellcarcinoma (Fig. 5a), squamous cell carcinoma (Fig. 5b),

squamous cell carcinoma in situ, adenocarcinoma, andadenocarcinoma in situ. In contrast, the nonneoplasticsquamous and glandular epithelia were either negative orshowed focal and weak cytoplasmic positivity. Strong anddiffuse nuclear staining for p53 protein was observed inone case of cervical squamous cell carcinoma (Fig. 5c) andin one case of squamous cell carcinoma in situ, which re-vealed TP53 missense mutation. Scattered tumor cells il-lustrated weak or moderate positivity for p53 with ratesranging from 1 to 60% (Fig. 5d), which meant wild typepattern of TP53. Three cases of composite tumors showednegative staining both in small cell carcinoma and inadenocarcinoma or squamous cell carcinoma in situ,which suggested TP53 nonsense mutation. The pure smallcell carcinoma of uterine cervix had similar mutation orwild type pattern of TP53 compared with compositetumor (P = 0.224), and there was no difference betweenTP53 mutation in small cell carcinoma and those in otherepithelial neoplasms of uterine cervix (P = 0.682).

DiscussionSmall cell carcinoma of the uterine cervix is a rare andhighly malignant tumor. The etiopathogenetic associ-ation between cervical small cell carcinoma and high-risk HPV infections has been well documented in somestudies [1, 2, 7]. Our study extend these findings bydemonstrating that all 30 cases of cervical small cell car-cinomas are related to high-risk HPV types 18 and 16,

Fig. 3 HPV DNA detection using in situ hybridization (× 400). a Small cell carcinoma illustrated punctate staining pattern with one navy-blue punctasignal in the nuclei. b Multiple small puncta inside the nuclei of small cell carcinoma represents low viral copy number of HPV DNA. c Diffuse smallparticles were observed in the nuclei of small cell carcinoma. d Squamous cell carcinoma displayed punctuate signal patterns. e Squamous cellcarcinoma illustrated episomal staining pattern. f Tumor cells of adenocarcinoma positive with one navy-blue puncta

Li et al. Diagnostic Pathology (2018) 13:31 Page 6 of 10

with predominant HPV 18 infection and multiple infec-tions with HPV 18 and 16. In addition, a high frequencyof cervical small cell carcinomas associated with otherepithelial tumors was reported. This may be related withthe common etiopathogenesis—HPV 18 and HPV 16.However, the prevalence of HPV 18 was different in cer-vical small cell carcinomas from squamous or glandularepithelial neoplasm.The mean age at diagnosis for women with cervical

small cell carcinoma was between 45 and 50 years [2],which is consistent with that of cervical squamous cellcarcinoma. In this study, the mean age of the patientswith cervical small cell carcinoma was 46.4 years. Thepatients had no specific clinical manifestations. Mostcases presented with abnormal vaginal bleeding or con-tact bleeding. Exophytic growth was not different fromother uterine cervical carcinomas. Of note, most of thepatients in our study were in the early stage, which dif-fers from previous studies [2, 4]. However, high rates oflymph node metastasis and poor prognosis were still ob-served in these stage I and II patients.Primary small cell carcinomas of uterine cervix often

coexist with squamous cell carcinomas or adenocarcin-omas. Wang et al. reported that in 22 cases of primarycervical small cell carcinomas, two exhibited concordanthigh-grade squamous intraepithelial neoplasia and adeno-carcinoma in situ [8]. In the study by Abeler et al., 12 ofthe 26 patients with cervical small cell carcinomas wereassociated with other forms of carcinoma, including squa-mous cell carcinoma (n = 6), adenocarcinoma (n = 5), andadenocarcinoma in situ (n = 1) [9]. Ishida et al. reported10 cases of cervical small cell carcinoma, 7 of which weremixed with adenocarcinoma and/or squamous cell carcin-oma, or cervical intraepithelial neoplasia [10]. Emerson etal. reported that in 19 cases of cervical small cell carcin-oma, 6 cases were associated with adenocarcinoma, andthree patients also had adenosquamous carcinoma, squa-mous cell carcinoma in situ, and adenocarcinoma [5]. Inthis study, 15 patients also displayed squamous or glandu-lar epithelial neoplasms, including squamous cell carcin-oma (n = 7), adenocarcinoma (n = 3), squamous cellcarcinoma in situ (n = 4), and adenocarcinoma in situ(n = 1). Our results demonstrate a high frequency of cer-vical small cell carcinomas associated with squamous orglandular epithelial tumors. They had similar age compos-ition and clinical manifestations as the patients of singlecervical tumor. Chan et al. found that pure, rather thanmixed histological pattern was a poor prognostic factorfor survival [11]. However, in our study, patients of cer-vical small cell carcinoma with and without other types ofepithelial neoplasms had similar prognosis, which was sig-nificantly worse than that of cervical squamous cell car-cinoma and adenocarcinoma [12]. Therefore, theprognosis of patients with composite cervical cancer was

Fig. 4 HPV typing using polymerase chain reaction-reverse dotblot hybridization

Li et al. Diagnostic Pathology (2018) 13:31 Page 7 of 10

determined by the composition of small cell carcinoma. Inaddition, in the case of biopsy specimens suspected of cer-vical epithelial disease, careful observation should be madeto avoid the omission of small cell carcinoma. In thisstudy, more lymphovascular invasion was observed in hys-terectomy and cervical conization specimens. Therefore,obtaining as many specimens as possible improves the ac-curacy of the diagnosis, and it will more accurately esti-mate the prognosis.High-risk HPV has been implicated in the carcinogenesis

of cervical small cell carcinomas [1, 7]. A meta-analysis in-cluding more than 30,000 invasive cervical cancers revealedthat HPV 16 (59%), 18 (13%), 58 (5%), 33 (5%), and 45 (4%)were the most prevalent subtypes in cervical squamous cellcarcinomas. HPV 18 (37%), 16 (36%), 45 (5%), 31 (2%), and33 (2%) were the most prevalent in cervical adenocarcin-omas [13]. Many studies have established that the preva-lence of different high-risk HPV types in cervical small cellcarcinoma ranged from 50 to 100%, and that HPV 18 maybe the most prevalent type [2]. Wang et al. reported that

HPV18 and 16 were detected in 77.3 and 18.2% cases ofcervical small cell carcinoma, respectively, and one case dis-played HPV 18 and 16 co-infection [8]. Research by Abeleret al. demonstrated that HPV-18 is predominant in puresmall cell carcinomas and in tumors with adenocarcinoma-tous areas, and that HPV-16 is found in pure small cell car-cinomas and in tumors with areas of squamous celldifferentiation [9]. In a study by Ishida et al., HPV 18 wasdetected in both small cell carcinomas and adenocarcino-matous components, and no other types of HPV were de-tected [10]. In our study, preponderant infection with HPV18 was found in 27 of 30 cases of cervical small cell carcin-oma. In the 15 cases of composite tumors, HPV 18 infec-tion was more common in small cell carcinoma than in anyother type of epithelial tumors. These findings are in linewith the notion that HPV 18 infection is involved in the de-velopment of cervical small cell carcinoma and HPV 16 in-fection promotes the occurrence of cervical squamous cellcarcinoma and adenocarcinoma.Multiple infections were observed in more than half of

the cervical small cell carcinomas in our study. Only onecase of multiple infections was found in 7 cases of squa-mous and glandular epithelial neoplasm. Multiple infec-tions of HPV 18 and 16, rather than pure HPV 18infection, might play an important role in the pathogen-esis of cervical small cell carcinoma. In previous studies,very few cases of multiple infections were reported [8].One explanation for this might be found in the differentdetection methods or the population differences. In theresearch by Zhou et al., using HPV DNA detection by

Table 2 HPV infection in small cell carcinoma and other typesof epithelial tumors

Small cellcarcinomas

Squamous or glandularepithelial neoplasms

P value

Cases (n) 22 7

HPV 18 19 2 0.008

HPV 16 14 6 0.382

HPV 18 + HPV 16 11 1 0.187

Fig. 5 Immunohistochemical expression of p16 and p53 (× 400). Strong and diffuse nuclear and cytoplasmic staining for p16 was noted incervical small cell carcinoma (a) and squamous cell carcinoma (b). c Nuclear staining for p53 was observed in cervical squamous cell carcinoma,wich revealed TP53 missense mutation. d Cervical small cell carcinoma illustrating scattered, weakly positive nuclear staining for p53 revealed wildtype pattern of TP53

Li et al. Diagnostic Pathology (2018) 13:31 Page 8 of 10

PCR-RDDH, multiple HPV DNA prevalence was 4.5% in2452 women whom volunteered for cervical cancerscreening in Shanghai, China [14]. A similar multiple in-fection prevalence rate was measured in 9012 womenwhom attended cervical cancer screening in Taihu RiverBasin, China [15]. In addition, reported HPV infectionrates are closely related to the detection methods used.For example, in our study, the ISH assessment for HPVinfection with low copy number was very difficult, dis-playing the lower sensitivity using ISH detection com-pared to PCR. Similar results were reported byMasumoto using ISH and direct sequencing of PCRproducts [16]. The prevalence of infection by multipleHPV genotypes was 20% in patients with cervical squa-mous cell carcinoma and adenocarcinoma [17]. One caseof multiple HPV infection was confirmed in seven casesof squamous or glandular epithelial neoplasms in ourstudy. Furthermore, the prevalence of multiple infectionsin small cell carcinoma was not significantly different inpure small cell carcinoma compared to the compositetumors, and multiple HPV infections did not affect theprognosis of these patients.Overexpression of p16 has been well documented in

high-risk HPV-related cervical squamous cell carcin-omas and adenocarcinomas, as well as their precursorlesions [18, 19]. Many recent studies have shown thatcervical small cell carcinoma also overexpressed p16protein [20, 21]. In this study, we detected simultaneousoverexpression of p16 in both small cell carcinoma andother types of epithelial tumors by immunohistochemis-try. Although the overexpression of p16 was correlatedwith HPV18 and 16 infections in both pure tumors andcomposite tumors, it does not confirm that high-riskHPV infections result in the overexpression of p16. In-deed, small cell carcinoma negative for HPV DNA in thelung, colorectum, bladder, and ovaries also overexpressp16 [8, 20]. In our study, the mutation or wild type pat-tern of TP53 in small cell carcinoma was not signifi-cantly different between pure and composite tumors.Three patients with composite tumors showed com-pletely negative p53 protein expression both in small cellcarcinoma and in other epithelial tumors. Furthermore,two cases revealed strong and diffuse positive expressionof p53 only in squamous cell carcinoma or squamouscell carcinoma in situ, while small cell carcinoma com-ponents of the same patients showed wild type pattern.These observations indicated TP53 mutation might in-volved in the occurrence of small cell carcinoma as insquamous cell carcinoma and adenocarcinoma of uterinecervix, but the pathogenesis of small cell carcinoma wasnot completely same to those of squamous cell carcin-oma or adenocarcinoma.Since only a single-center retrospective case series was

studied, and limited cases were employed, this study

should be repeated on a larger scale to confirm our find-ings. Furthermore, some patients did not complete theirfollow-up, thus the comprehensive and effective datawere not obtained.

ConclusionsIn summary, our study demonstrated that cervical smallcell carcinomas closely correlate with HPV18 and HPV16 infections. The patients of cervical small cell carcin-omas with multiple infection of high-risk HPV may alsopromote the development of squamous or glandular epi-thelial neoplasms. In patients with composite cervicalneoplasms, multiple infections of HPV 18 and 16 wereinvolved in the development of cervical small cell carcin-oma, while the occurrence of cervical squamous cell car-cinoma or adenocarcinoma was closely related to HPV16 infection. Multiple high-risk HPV infection was notrelated to tumor stage, size, lymphovascular invasion,lymph node metastasis, or prognosis in cervical smallcell carcinoma. Similar HPV DNA genotypes, clinico-pathological features, and prognosis were observed bothin patients with pure small cell carcinomas and thosewith composite cervical tumors. The small cell carcin-oma determined the prognosis of patients with cervicalcomposite tumors. Because of the frequent presence ofco-existing tumors, it is important to carefully examinethe cervical biopsy specimens in surgical pathologicalexamination.

AbbreviationsFIGO: International Federation of Gynecology and Obstetrics; HPV: Humanpapillomavirus; ISH: In Situ Hybridization; PCR: Polymerase chain reaction;RDDH: Reverse dot blot hybridization

AcknowledgementsWe are grateful to the Department of Pathology, Xijing Hospital for excellentworking conditions and for providing access to archival materials.

FundingThe study was funded by grants from the National Natural ScienceFoundation of China (81272651 and 81570180 to Zhe Wang).

Availability of data and materialsThe data are available upon request on the following e-mail address:lipeifeng00@hotmail.com.

Authors’ contributionsZW study design, PL, JM, XZ data acquisition, YL, DZ ISH analysis and PCR,PL, JM, XZ, YG data analysis, PL, JM, XL statistical analysis PL, JM manuscriptpreparation and editing. All authors read and approved the final manuscript.

Ethics approval and consent to participateThis study received approval from the Ethics Committee of the XijingHospital [KY20173326–1]. Additional informed consent was not required bythe ethics committee for this retrospective study.

Competing interestsThe authors declare that they have no competing interests.

Publisher’s NoteSpringer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Li et al. Diagnostic Pathology (2018) 13:31 Page 9 of 10

Received: 5 February 2018 Accepted: 6 May 2018

References1. Small W Jr, Bacon MA, Bajaj A, Chuang LT, Fisher BJ, Harkenrider

MM, Jhingran A, Kitchener HC, Mileshkin LR, Viswanathan AN,Gaffney DK. Cervical cancer: a global health crisis. Cancer. 2017;123:2404–12.

2. Atienza-Amores M, Guerini-Rocco E, Soslow RA, Park KJ, Weigelt B. Small cellcarcinoma of the gynecologic tract: a multifaceted spectrum of lesions.Gynecol Oncol. 2014;134:410–8.

3. Satoh T, Takei Y, Treilleux I. Gynecologic Cancer InterGroup (GCIG)consensus review for small cell carcinoma of the cervix. Int J GynecolCancer. 2014;24(3):S102–8.

4. Dores GM, Qubaiah O, Mody A, Ghabach B, Devesa SS. A population-basedstudy of incidence and patient survival of small cell carcinoma in theUnited States, 1992-2010. BMC Cancer. 2015;15:185.

5. Emerson RE, Michael H, Wang M, Zhang S, Roth LM, Cheng L. Cervicalcarcinomas with neuroendocrine differentiation: a report of 28 cases withImmunohistochemical analysis and molecular genetic evidence of commonclonal origin with coexisting squamous and adenocarcinomas. Int J GynecolPathol. 2016;35:372–84.

6. Carow K, Read C, Hafner N, Runnebaum IB, Corner A, Dürst M. Acomparative study of digital PCR and real-time qPCR for the detection andquantification of HPV mRNA in sentinel lymph nodes of cervical cancerpatients. BMC Res Notes. 2017;10:532.

7. Gadducci A, Carinelli S, Aletti G. Neuroendrocrine tumors of the uterinecervix: a therapeutic challenge for gynecologic oncologists. Gynecol Oncol.2017;144:637–46.

8. Wang HL, Lu DW. Detection of human papillomavirus DNA and expressionof p16, Rb, and p53 proteins in small cell carcinomas of the uterine cervix.Am J Surg Pathol. 2004;28:901–8.

9. Abeler VM, Holm R, Nesland JM, Kjørstad KE. Small cell carcinoma ofthe cervix. A clinicopathologic study of 26 patients. Cancer.1994;73:672–7.

10. Ishida GM, Kato N, Hayasaka T, Saito M, Kobayashi H, Katayama Y, Sasou S,Yaegashi N, Kurachi H, Motoyama T. Small cell neuroendocrine carcinomasof the uterine cervix: a histological, immunohistochemical, and moleculargenetic study. Int J Gynecol Pathol. 2004;23:366–72.

11. Chan JK, Loizzi V, Burger RA, Rutgers J, Monk BJ. Prognostic factors inneuroendocrine small cell cervical carcinoma: a multivariate analysis.Cancer. 2003;97:568–74.

12. Zhou J, Wu SG, Sun JY, Tang LY, Lin HX, Li FY, Chen QH, Jin X, He ZY.Clinicopathological features of small cell carcinoma of the uterine cervixin the surveillance, epidemiology, and end results database. Oncotarget.2017;8:40425–33.

13. Li N, Franceschi S, Howell-Jones R, Snijders PJ, Clifford G. Humanpapillomavirus type distribution in 30,848 invasive cervical cancersworldwide: variation by geographical region, histological type and year ofpublication. Int J Cancer. 2011;128:927–35.

14. Zhou XH, Shi YF, Wang LJ, Liu M, Li F. Distribution characteristics of humanpapillomavirus infection: a study based on data from physical examination.Asian Pac J Cancer Prev. 2017;18:1875–9.

15. Lu JF, Shen GR, Li Q. Genotype distribution characteristics of multiplehuman papillomavirus in women from the Taihu River basin, on the coastof eastern China. BMC Infect Dis. 2017;17:226.

16. Masumoto N, Fujii T, Ishikawa M, Saito M, Iwata T, Fukuchi T, Susumu N,Mukai M, Kubushiro K, Tsukazaki K, Nozawa S. P16 overexpression andhuman papillomavirus infection in small cell carcinoma of the uterinecervix. Hum Pathol. 2003;34:778–83.

17. Mazarico E, Gomez-Roig MD, Minano J, Cortes L, Gonzalez-Bosquet E.Relationship of human papilloma virus multiple genotype infection withpatient’s age and type of cervical lesion. Eur J Gynaecol Oncol. 2014;35:378–81.

18. Ahmad A, Raish M, Shahid M. Batra S4, Batra V5, Husain S. Thesynergic effect of HPV infection and epigenetic anomaly of the p16gene in the development of cervical cancer. Cancer Biomark. 2017;19:375–81.

19. Togami S, Sasajima Y, Kasamatsu T, Oda-Otomo R, Okada S, Ishikawa M,Ikeda S, Kato T, Tsuda H. Immunophenotype and human papillomavirus

status of serous adenocarcinoma of the uterine cervix. Pathol Oncol Res.2015;21:487–94.

20. Carlson JW, Nucci MR, Brodsky J, Crum CP, Hirsch MS. Biomarker-assisteddiagnosis of ovarian, cervical and pulmonary small cell carcinomas: the roleof TTF-1, WT-1 and HPV analysis. Histopathology. 2007;51:305–12.

21. Horn LC, Hauptmann S, Leo C. pRb, p16, and cyclin D1 in small cellcarcinoma of the uterine cervix. Int J Gynecol Pathol. 2007;26:269.

Li et al. Diagnostic Pathology (2018) 13:31 Page 10 of 10